Cassandra CPU perfomance

["D. Salvatore" <dd...@gmail.com>]

Hi,

I deployed a Cassandra 2.2 ring composed by 4 nodes in the cloud with 8
vCPU and 8GB of ram. I am running some tests now with cassandra-stress and
YCSB tools to test its performance. I am mainly interested in read requests
with a small amount of write requests (95%/5%).

Running the experiments, I noticed that even setting a high number of
threads (or clients) the CPU (and disk) does not saturate, but still always
around the 60% of utilisation.

I am trying to figure out where is the bottleneck in my system. From the
hardware point of view it seems all ok to me.

I also looked into the Cassandra configuration file to see if there are
some tuning parameters to increase the system throughput. I increase the
value of concurrent_read/write parameter, but it doesn't increase the
performance. The log file also does not contain any warning.

What it could be that is limiting my system?

Thanks

RE: Cassandra CPU perfomance

[DE VITO Dominique <do...@thalesgroup.com>]

Hi,

A hint : depending on your data set size + your request rate per second, 8 GB of RAM may be too low.
And then, CPU might be high due to too frequent GC.

More RAM may bring:

·         More space for OS FS to cache the SSTable files in memory.

·         A greater heap size, and then, less frequent object promotion (young=>old space) and then, less major GC.


Dominique



[@@ THALES GROUP INTERNAL @@]

De : D. Salvatore [mailto:dd.salvatore@gmail.com]
Envoyé : mercredi 4 janvier 2017 16:28
À : user@cassandra.apache.org
Objet : Cassandra CPU perfomance


Hi,

I deployed a Cassandra 2.2 ring composed by 4 nodes in the cloud with 8 vCPU and 8GB of ram. I am running some tests now with cassandra-stress and YCSB tools to test its performance. I am mainly interested in read requests with a small amount of write requests (95%/5%).

Running the experiments, I noticed that even setting a high number of threads (or clients) the CPU (and disk) does not saturate, but still always around the 60% of utilisation.

I am trying to figure out where is the bottleneck in my system. From the hardware point of view it seems all ok to me.

I also looked into the Cassandra configuration file to see if there are some tuning parameters to increase the system throughput. I increase the value of concurrent_read/write parameter, but it doesn't increase the performance. The log file also does not contain any warning.

What it could be that is limiting my system?

Thanks